Skip to content
Back to search

reference-video-director

Creates production-ready reference-driven video prompts from character, outfit, scene, relationship, story, and duration inputs. Use for requests such as 情侣短剧、情侣互动、男友视角、女友视角、情侣POV、恋爱POV、参考图视频、15秒/30秒剧情、couple POV, boyfriend POV, girlfriend POV, romantic short scene, reference video prompt, RISE火焰视频、点烟抛火柴连续镜头、燃烧文字、RISE fire video, burning-letter reveal, or when a rough scene needs natural dialogue, micro-expressions, continuity, and model-specific optimization.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/sgyno09-source--reference-video-director/badge.svg)](/listings/sgyno09-source--reference-video-director)

HTML

<a href="/listings/sgyno09-source--reference-video-director"><img src="/api/listings/sgyno09-source--reference-video-director/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

这个仓库本身就是一个 Skill 目录,核心入口文件是:

SKILL.md

兼容 Agent Skills / Codex Skill 风格的客户端,可直接使用仓库目录作为 Skill 来源。

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 496 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. ---
  2. name: reference-video-director
  3. description: Creates production-ready reference-driven video prompts from character, outfit, scene, relationship, story, and duration inputs. Use for requests such as 情侣短剧、情侣互动、男友视角、女友视角、情侣POV、恋爱POV、参考图视频、15秒/30秒剧情、couple POV, boyfriend POV, girlfriend POV, romantic short scene, reference video prompt, RISE火焰视频、点烟抛火柴连续镜头、燃烧文字、RISE fire video, burning-letter reveal, or when a rough scene needs natural dialogue, micro-expressions, continuity, and model-specific optimization.
  4. ---
  5. # Reference Video Director
  6. ## Specialized fire-video workflow
  7. For RISE fire videos, lighting a cigarette and throwing a match, burning-letter reveals, or variants of this sequence, read [RISE Fire Video Prompt](skills/create-rise-fire-video/SKILL.md) and its [30-second baseline](skills/create-rise-fire-video/references/rise-30s.md). Follow that workflow for the full prompt and validation.
  8. In this mode, use its single continuous dolly-back shot, calm performance, mouth-held cigarette, one-match causality, delayed leg crossing, and unobstructed final composition. Do not apply the couple relationship defaults, changing eye-line requirements, recommended multi-shot counts, or couple dialogue/story engine below. Preserve the shared identity and outfit continuity rules. Treat RISE as scene geometry, not a prohibited text overlay. User-requested changes remain authoritative.
  9. For other scenes, continue with the general workflow below. Do not introduce smoking, fire, or burning letters into unrelated couple scenes.
  10. ## Purpose
  11. Turn a user's reference images and rough story idea into an executable short-form video prompt with stable identity, believable acting, natural dialogue, continuous blocking, realistic POV cinematography, and model-aware constraints.
  12. Default to the user's current language. If the user writes Chinese, answer in Chinese. If the user writes English, answer in English. If the user explicitly requests another language, follow that request.
  13. ## Trigger logic
  14. Activate when the user asks for one or more of the following:
  15. - reference-image-driven video prompting
  16. - couple / relationship POV
  17. - boyfriend POV or girlfriend POV
  18. - romantic daily-life short scenes
  19. - natural couple dialogue
  20. - fake anger, teasing, awkward comforting, playful banter
  21. - 10s / 15s / 30s / 60s narrative video prompts
  22. - timeline-based prompt writing
  23. - character identity locking
  24. - outfit or scene continuity
  25. - optimization for Seedance, Hailuo, Kling, Veo, Sora, or comparable video generators
  26. Common Chinese triggers include:
  27. `情侣短剧` `情侣互动` `恋爱短剧` `男友视角` `女友视角` `情侣POV` `恋爱POV` `女友假生气` `男友哄女友` `情侣聊天` `打情骂俏` `甜蜜互动` `参考图视频` `视频剧情提示词`.
  28. Common English triggers include:
  29. `couple POV` `boyfriend POV` `girlfriend POV` `romantic short film` `couple interaction` `relationship POV` `romantic dialogue` `realistic couple video` `reference video prompt`.
  30. Do not activate for unrelated general writing or purely technical video encoding tasks.
  31. ## Inputs
  32. Infer as many of these as available. Do not force the user to fill every field.
  33. - `{{Mixed 1}}`: primary character reference, usually the female lead
  34. - `{{Mixed 2}}`: outfit reference
  35. - `{{Mixed 3}}`: optional scene / environment reference
  36. - relationship
  37. - personality
  38. - story beat
  39. - target duration
  40. - target model
  41. - aspect ratio
  42. - audio requirements
  43. - camera POV
  44. - continuity constraints
  45. When information is missing, make conservative defaults that preserve the user's concept.
  46. ## Character consistency
  47. When a character reference is provided, treat it as the identity source of truth.
  48. Lock:
  49. - face shape
  50. - facial proportions
  51. - eyes
  52. - nose
  53. - lips
  54. - eyebrows
  55. - hairstyle
  56. - hair color
  57. - skin tone
  58. - perceived adult age
  59. - body proportions
  60. - overall recognizability
  61. Do not redesign the face unless the user explicitly asks.
  62. Prevent:
  63. - face swapping
  64. - identity drift
  65. - sudden makeup changes
  66. - hairstyle changes
  67. - age changes
  68. - unexplained body-proportion changes
  69. If the user provides multiple character references, clearly map each reference to one character.
  70. ## Outfit consistency
  71. When an outfit reference is supplied, lock visible clothing attributes:
  72. - garment type
  73. - color
  74. - material
  75. - straps
  76. - sleeves
  77. - neckline
  78. - lower garment
  79. - footwear
  80. - accessories
  81. Do not allow unexplained wardrobe changes within a continuous scene.
  82. ## Scene consistency
  83. When a scene reference is supplied, preserve the major spatial anchors:
  84. - room shape
  85. - bed / sofa / table
  86. - windows
  87. - curtains
  88. - mirrors
  89. - lamps
  90. - doors
  91. - major furniture placement
  92. - lighting direction
  93. - time of day
  94. Do not turn an ordinary room into a luxury hotel unless requested.
  95. ## Default relationship tone
  96. If the user does not specify otherwise, assume the characters are adult partners in a stable, familiar relationship.
  97. Favor:
  98. - familiarity
  99. - teasing
  100. - casual physical proximity
  101. - unfinished sentences
  102. - little pauses
  103. - playful complaints
  104. - shared jokes
  105. - believable reactions
  106. Avoid:
  107. - forced flirting
  108. - domineering romance tropes
  109. - constant intense eye contact
  110. - melodramatic confession language
  111. - scripted idol-drama behavior
  112. ## Dialogue rules
  113. Dialogue should be short, speakable, and slightly imperfect.
  114. Prefer:
  115. - short lines
  116. - interruptions
  117. - half-finished thoughts
  118. - small pauses
  119. - brief repetition
  120. - understated replies
  121. - reaction before response
  122. Avoid long polished speeches.
  123. A useful principle:
  124. `subtext > literal romantic wording`
  125. If the dialogue does not comfortably fit the duration, remove lines rather than making the characters speak unnaturally fast.
  126. ## Micro-expression rules
  127. Translate abstract emotions into visible behavior.
  128. Do not write only:
  129. `she gets angry` or `she becomes happy`.
  130. Instead describe sequences such as:
  131. - smile remains for a beat
  132. - she realizes what he said
  133. - the smile fades slightly
  134. - eyes narrow a little
  135. - lips press together
  136. - one corner of the mouth almost rises
  137. - she looks away
  138. - secretly glances back
  139. - tries not to laugh
  140. - lowers her head
  141. - shoulders move with suppressed laughter
  142. - finally breaks into a genuine smile
  143. For playful fake anger, emphasize the "trying not to laugh" progression.
  144. Avoid influencer-style exaggerated faces.
  145. ## Eye-line rules
  146. Do not make the lead stare into the lens continuously.
  147. Use changing eye behavior:
  148. `look at partner → look away → lower gaze → glance back → pretend to stay annoyed → suppress smile → re-establish eye contact`
  149. The eye-line must support the emotional beat.
  150. ## POV and camera
  151. Default to realistic close-distance relationship POV when appropriate.
  152. For boyfriend POV:
  153. - the boyfriend's face must remain off-screen
  154. - a hand, forearm, sleeve edge, or blurred shoulder may appear
  155. - never reveal the full male body unless requested
  156. - treat the female lead looking into camera as looking at her partner
  157. Use mild handheld breathing motion.
  158. Avoid:
  159. - aggressive push-ins
  160. - rapid zooms
  161. - unnecessary orbit shots
  162. - music-video camera movement
  163. - a cut for every dialogue line
  164. Recommended shot counts:
  165. - 10s: 1–2 shots
  166. - 15s: 1–3 shots
  167. - 30s: 3–4 longer shots
  168. - 60s: usually split into two 30s parts if the target model has a 30s generation limit
  169. ## Mirror rule
  170. If a mirror exists, explicitly prevent accidental reflections of:
  171. - partner's face
  172. - camera operator
  173. - camera / phone / rig
  174. - crew
  175. - third person
  176. Mirror geometry should remain consistent with the room.
  177. ## Physical interaction
  178. Use only a few simple, readable actions.
  179. Examples:
  180. - light cheek pinch
  181. - move a strand of hair
  182. - brief hand touch
  183. - push a hand away
  184. - tug a sleeve
  185. - touch nose tip
  186. - grab a pillow
  187. - block the camera
  188. - lean closer
  189. - short hug
  190. Interactions should be brief, natural, and physically plausible.
  191. For hand contact, include:
  192. - correct finger count
  193. - stable hand anatomy
  194. - no fusion
  195. - no extra hands
  196. - no penetration
  197. - real contact
  198. - natural entry and exit from frame
  199. - no sudden disappearance
  200. ## Story engine
  201. For light couple scenes, prefer:
  202. `normal conversation → tiny misunderstanding / tease → small emotional reaction → awkward attempt to fix it → unintended funny line → suppressed laughter → small reversal → genuine reaction`
  203. Possible everyday hooks:
  204. - one person talks too much
  205. - blanket stealing
  206. - room temperature
  207. - forgotten reply
  208. - ugly photo
  209. - snacks
  210. - pillow fight
  211. - wrong nickname
  212. - gaming result
  213. - who turns off the light
  214. - who gets water
  215. - messy hair
  216. - one person refusing to admit they are smiling
  217. Avoid major betrayal, breakup, death, or third-party drama unless explicitly requested.
  218. ## Duration logic
  219. ### Under 10 seconds
  220. Use:
  221. - one setup
  222. - one interaction
  223. - one payoff
  224. ### 15 seconds
  225. Typical structure:
  226. - 00:00–00:04 setup
  227. - 00:04–00:09 tiny conflict
  228. - 00:09–00:13 interaction
  229. - 00:13–00:15 payoff
  230. Usually 4–7 short dialogue lines maximum.
  231. ### 30 seconds
  232. Typical structure:
  233. - 00:00–00:06 natural setup
  234. - 00:06–00:13 conflict develops
  235. - 00:13–00:19 physical interaction
  236. - 00:19–00:25 comic escalation
  237. - 00:25–00:30 reversal / laugh / ending beat
  238. Usually 8–15 short dialogue lines, depending on pacing.
  239. ### 60 seconds
  240. If the chosen model is practically limited to ~30 seconds per generation, produce:
  241. - Part 1: 00:00–00:30
  242. - Part 2: 00:30–01:00
  243. Each part should have a mini-payoff and a clean continuity hook.
  244. ## Audio
  245. Unless requested otherwise for natural daily-life scenes:
  246. - no background music
  247. - no subtitles
  248. - no text overlays
  249. - no narration
  250. Keep realistic diegetic sound:
  251. - room tone
  252. - air conditioner
  253. - bedding friction
  254. - clothing friction
  255. - hair movement
  256. - breathing
  257. - suppressed laughter
  258. - natural speech
  259. Leave small pauses between lines.
  260. ## Model adaptation
  261. ### Seedance
  262. Prioritize:
  263. - clean continuous actions
  264. - explicit time ranges
  265. - moderate dialogue density
  266. - limited complex hand choreography
  267. - clear cause-and-effect blocking
  268. ### Hailuo
  269. Prioritize:
  270. - strong identity lock
  271. - explicit action descriptions
  272. - one main action at a time
  273. - clear continuity reminders
  274. ### Kling
  275. Prioritize:
  276. - spatial clarity
  277. - camera-to-subject relation
  278. - consistent blocking
  279. - slightly more cinematic camera language when useful
  280. ### Veo / Sora
  281. May include richer:
  282. - sound cues
  283. - environmental details
  284. - lens / camera behavior
  285. - dialogue timing
  286. - subtle performance direction
  287. Do not add technical camera jargon unless it improves execution.
  288. ## Output format
  289. Produce a clean, directly copyable prompt with these sections when relevant:
  290. 1. Character Consistency
  291. 2. Outfit Consistency
  292. 3. Core Setting
  293. 4. Scene
  294. 5. Camera / POV
  295. 6. Timeline
  296. 7. Performance Notes
  297. 8. Male / Off-screen Partner Performance
  298. 9. Continuity
  299. 10. Sound Design
  300. 11. Negative Constraints
  301. 12. Model-specific notes
  302. For each timeline block, include as needed:
  303. - action
  304. - micro-expression
  305. - eye-line
  306. - dialogue
  307. - pause
  308. - camera behavior
  309. Do not mechanically include empty sections.
  310. ## Editing an existing user prompt
  311. If the user already wrote a detailed prompt, preserve the core story.
  312. Apply this order:
  313. `preserve story → reduce redundant dialogue → add pauses → add eye-line changes → add micro-expressions → fix physical continuity → reduce melodrama → improve relationship familiarity → simplify actions that are difficult for video models`
  314. Do not overwrite distinctive user ideas unless they conflict with the user's stated constraints.
  315. ## Negative constraints
  316. Use relevant items, not necessarily every line:
  317. - No face identity drift.
  318. - No face swapping.
  319. - No hairstyle changes.
  320. - No outfit changes.
  321. - No sudden makeup changes.
  322. - No unexplained age changes.
  323. - No extra people.
  324. - No visible camera operator.
  325. - No visible recording equipment.
  326. - No partner face in mirrors when POV requires them to remain unseen.
  327. - No extra hands.
  328. - No malformed fingers.
  329. - No hand penetration.
  330. - No disappearing hands.
  331. - No exaggerated facial acting.
  332. - No influencer-style acting.
  333. - No idol-drama acting.
  334. - No melodramatic romance.
  335. - No excessive blinking.
  336. - No fixed staring.
  337. - No subtitles unless requested.
  338. - No text unless requested.
  339. - No background music unless requested.
  340. - No random scene changes.
  341. - No jumpy continuity.
  342. ## Final quality check
  343. Before returning the prompt, verify:
  344. - one stable identity per character
  345. - adult characters when romance is involved
  346. - outfit continuity
  347. - location continuity
  348. - dialogue fits duration
  349. - pauses remain possible
  350. - POV rules are respected
  351. - mirror reflections are safe
  352. - hand actions are feasible
  353. - no unnecessary extra characters
  354. - emotions are expressed through visible micro-actions
  355. - ending has a readable final beat
  356. The final result should feel less like a scripted romance scene and more like a believable moment that happened to be recorded.
README.md · README · 338 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Reference Video Director
  2. <p align="center">
  3. <a href="./README.md"><img alt="中文" src="https://img.shields.io/badge/语言-中文-111111"></a>
  4. <a href="./README_EN.md"><img alt="English" src="https://img.shields.io/badge/Language-English-111111"></a>
  5. </p>
  6. **Reference Video Director** 是一个面向参考图驱动视频创作的 Agent Skill。
  7. 它会根据角色参考图、服装参考图、场景参考图、人物关系、剧情想法和目标时长,生成更适合视频模型直接执行的完整提示词,重点处理人物一致性、情侣 POV、自然对白、微表情、手部互动、镜头连续性与真实生活感。
  8. # 🚀 新手 3 分钟上手
  9. 完全没用过 Skill 也没关系,按下面做就行。
  10. ## 第一步:下载
  11. ### 方法 A:直接下载 ZIP(最适合新手)
  12. 打开本仓库后,点击:
  13. ```text
  14. Code → Download ZIP
  15. ```
  16. 下载完成后解压,得到 `reference-video-director` 文件夹。
  17. **不要只下载 `SKILL.md`,建议保留整个文件夹结构。**
  18. ### 方法 B:Git Clone
  19. 会用 Git 的话可以直接:
  20. ```bash
  21. git clone https://github.com/sgyno09-source/reference-video-director.git
  22. ```
  23. ## 第二步:安装 / 导入
  24. 这个仓库本身就是一个 Skill 目录,核心入口是:
  25. ```text
  26. SKILL.md
  27. ```
  28. 如果你的 Agent、CLI、Codex 或其他客户端支持 Skills:
  29. 1. 找到它的 Skills / Agent Skills / Custom Skills 导入入口。
  30. 2. 选择刚才下载并解压的 `reference-video-director` 文件夹。
  31. 3. 保持 `SKILL.md` 和 `references/` 等文件原有位置不变。
  32. 4. 导入完成后重新打开会话或刷新 Skill 列表。
  33. 不同客户端的安装入口可能不同,但原则一样:**导入整个 Skill 文件夹,而不是把里面的文字复制出来。**
  34. ## 第三步:开始使用
  35. 安装后不需要背复杂命令。
  36. 上传参考图,然后直接告诉它:
  37. ```text
  38. 参考这张女生图
  39. 做一个30秒情侣POV
  40. 保持同一个人物和发型
  41. 服装不要变化
  42. 场景是晚上卧室
  43. 女生性格俏皮爱笑
  44. 男友不露脸
  45. 剧情:
  46. 男友说错一句话
  47. 女生先愣一下
  48. 然后假装生气
  49. 最后又被逗笑
  50. 模型:Seedance
  51. 对白自然一点
  52. 不要字幕
  53. ```
  54. Skill 会帮你自动补全:
  55. ```text
  56. 参考图绑定
  57. 人物与服装一致性
  58. 场景与光线
  59. Camera / POV
  60. 分秒时间轴
  61. 动作与微表情
  62. 对白与停顿
  63. 手部互动
  64. 声音设计
  65. 连续性控制
  66. 负面约束
  67. 模型适配
  68. ```
  69. 你负责给**参考图 + 大概剧情**,剩下的交给 Skill。
  70. ## 多张参考图怎么写
  71. 如果同时上传人物、服装、场景参考图,最好明确告诉它每张图负责什么:
  72. ```text
  73. Mixed 1:女主人物参考
  74. Mixed 2:服装参考
  75. Mixed 3:卧室场景参考
  76. 使用 Mixed 1 作为唯一女主角
  77. 服装严格参考 Mixed 2
  78. 场景结构参考 Mixed 3
  79. 生成30秒情侣POV
  80. ```
  81. 这样可以减少换脸、换衣服、场景漂移等问题。
  82. ## 只会一句剧情也可以
  83. 不需要自己会写分镜。
  84. 例如直接说:
  85. ```text
  86. 参考这个女生
  87. 做一个30秒视频
  88. 男朋友说错一句话
  89. 她从开心到愣住
  90. 再到假装生气
  91. 最后被逗笑
  92. 帮我完整设计
  93. ```
  94. Skill 会自动把一句话扩展成可执行的视频提示词。
  95. ## 做 15 秒 / 30 秒视频
  96. 直接告诉它时长即可:
  97. ```text
  98. 做15秒,节奏快一点,只保留一个转折
  99. ```
  100. 或者:
  101. ```text
  102. 做30秒,保留起因、假生气、哄人和最后笑出来
  103. ```
  104. Skill 会根据时长自动控制对白和动作密度,不会把一分钟剧情硬塞进 30 秒。
  105. ## 做两个连续的 30 秒
  106. 如果模型单次只能生成 30 秒,可以这样说:
  107. ```text
  108. 这段剧情一共60秒
  109. 拆成两个30秒
  110. 第二段必须接第一段最后一个动作
  111. 人物、服装、发型、场景、灯光和情绪状态保持一致
  112. ```
  113. Skill 会为第二段补上上一段结束时的人物位置、动作和状态,方便连续生成。
  114. ## POV / 男友不露脸怎么写
  115. 例如:
  116. ```text
  117. 男友使用第一人称POV
  118. 全程不露脸
  119. 只允许手、前臂、肩部或画外音进入画面
  120. 镜子里不要出现男友和摄影设备
  121. ```
  122. 这类空间关系、镜面和手部互动也是本 Skill 重点处理的部分。
  123. ## 指定视频模型
  124. 可以直接写模型名称:
  125. ```text
  126. 模型:Seedance
  127. ```
  128. ```text
  129. 模型:Hailuo
  130. ```
  131. ```text
  132. 模型:Sora
  133. ```
  134. Skill 会根据模型调整动作密度、时间轴、镜头与连续性描述。没有指定模型时,会输出通用版本。
  135. ## 一个推荐的新手模板
  136. 不会写需求时,复制下面这段改内容就行:
  137. ```text
  138. 参考图:
  139. 图1是人物
  140. 图2是服装
  141. 图3是场景
  142. 时长:30秒
  143. 模型:Seedance
  144. 画幅:9:16
  145. 人物关系:情侣
  146. 女生性格:俏皮、爱笑、情绪写在脸上
  147. 男友:不露脸,第一人称POV
  148. 剧情:
  149. 这里写一句你想发生的事情
  150. 要求:
  151. 保持同一人物
  152. 保持同一服装
  153. 对白自然
  154. 加强眼神和微表情
  155. 动作不要太密
  156. 不要字幕
  157. 不要突然换场景
  158. ```
  159. 把最后的剧情换成自己的想法,就可以开始用了。
  160. # 适合什么
  161. - 情侣 POV / 恋爱短剧
  162. - 女友假生气 / 男友哄人
  163. - 居家、卧室、旅行、餐厅等生活化场景
  164. - 参考图角色一致性
  165. - 15 秒 / 30 秒短剧情
  166. - Seedance、Hailuo、Kling、Veo、Sora 等视频模型
  167. - 将粗略剧情改造成时间轴 Prompt
  168. - 将“太像演戏”的对白改得更自然
  169. ## 核心原则
  170. **真实感 > 戏剧性**
  171. **人物反应 > 漂亮台词**
  172. **情侣默契 > 强行暧昧**
  173. **微表情 > 大动作**
  174. **停顿 > 连续念台词**
  175. **小事情 > 大冲突**
  176. ## 输入示例
  177. ```text
  178. 使用 {{Mixed 1}} 作为唯一女主角参考,
  179. 服装参考 {{Mixed 2}}。
  180. 做一个 30 秒男友 POV 情侣短剧:
  181. 女友本来在聊天,男友说“你今天话好多”,
  182. 女友开始假生气,最后被男友一句话逗笑。
  183. 模型:Seedance
  184. ```
  185. Skill 会自动完成:
  186. - Character Consistency
  187. - Outfit Consistency
  188. - Core Setting
  189. - Scene
  190. - Camera / POV
  191. - Timeline
  192. - Dialogue
  193. - Micro-expressions
  194. - Physical Interaction
  195. - Continuity
  196. - Sound Design
  197. - Negative Constraints
  198. - Model Adaptation
  199. ## 安装
  200. 这个仓库本身就是一个 Skill 目录,核心入口文件是:
  201. ```text
  202. SKILL.md
  203. ```
  204. 兼容 Agent Skills / Codex Skill 风格的客户端,可直接使用仓库目录作为 Skill 来源。
  205. ## 目录
  206. ```text
  207. reference-video-director/
  208. ├── SKILL.md
  209. ├── README.md
  210. ├── README_EN.md
  211. └── references/
  212. ├── zh-CN.md
  213. └── en.md
  214. ```
  215. ## 语言
  216. Skill 会自动跟随用户当前使用的语言。
  217. - 中文输入 → 中文输出
  218. - English input → English output
  219. - 用户明确指定语言 → 按指定语言输出
  220. ## RISE 火焰电影提示词
  221. 仓库新增独立技能 [RISE 火焰电影提示词](skills/create-rise-fire-video/SKILL.md),可单独导入 `skills/create-rise-fire-video/` 整个目录,也可以通过主技能自动调用。
  222. 默认方案为30秒、16:9连续镜头:点烟、吐烟、抛出同一根火柴、落地后火势沿RISE路径扩展、镜头持续后拉、22秒后翘二郎腿,最终完整人物、椅子、双脚与RISE互不遮挡。支持换人物、换文字和调整时长。
  223. 人物参考图需由使用者提供。技能输出视频提示词,不会自行生成视频。
  224. 调用示例:
  225. ```text
  226. 使用 create-rise-fire-video
  227. 图1作为唯一人物参考
  228. 做30秒RISE火焰视频完整提示词
  229. 保持全身、椅子和字母完整无遮挡
  230. ```
  231. 完整时间轴与约束见 [30秒方案](skills/create-rise-fire-video/references/rise-30s.md)。
  232. ---
  233. Made for reference-driven video prompt workflows.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.